Transcript (MBSE) Model Based System Engineering Integration

INCOSE (MBSE)
Model Based System Engineering
Integration and Verification Scenario
Ron Williamson, PhD
Raytheon
[email protected]
Jan 21-22, 2012
INCOSE IW12 MBSE Workshop
INCOSE IW12 MBSE Workshop
15
Integration & Verification Scenario
• Concept of Operations
– Word document and Scenarios Model (Behavior/Structure)
• Requirements & Architecture
– Word Document and Requirement Objects, Architecture Framework Languages
• Design and Implementation
– Diagrams and Word document à SW and HW Models (Form & Function)
• Test
– Plans and Test Models
• Integration
– Model Integration
• Verification
– Model Verification
• Operation and Maintenance
– Operational Models
INCOSE IW12 MBSE Workshop
17
Integration & Verification Scenario
Project Modeling
Concept of
Operations
Model
Project
Definition
Operations &
Maintenance
Operational Parameters
Model
Synch
Model Verify & Validate
Model vs Implementation
Constraints & Viewpoints
Requirements
& Architecture
Model Verify & Validate
Models
Model
Synch
Structure & Behaviors
Detailed Design
Model
Models
Model Verify & Validate
System Verif &
Validation
Model vs Impl.
Project
Integration, Test Test &
& Verification Integration
Synch
Model vs Impl.
Implementation
Time
INCOSE IW12 MBSE Workshop
18
Integration & Test Scenario
• Problem Description
• Challenges: It's all about better communication!!
• Use Cases
• Opportunities: How can MBSE help?
• Experience (lessons learned)
• Methodologies (Process, Methods, Tools)
Challenge Summary
• Communication gaps and/or mis-understanding across teams, disciplines and
phases
• Funding early involvement of integration and verification planners and
designers
• IV&V teams are not working in the model
• Lack of consistency across lifecycle team
• IV&V ends up fixing problems from early phases, increasing risk, cost and
schedule
• No benchmarking has been performed to help gather metrics
• Lack of knowledge transfer from integrator/tester to the knowledge base
(manually or via tool)
Use Case Summary
•
Choose development model
•
Introduce inter disciplines early
•
Share models across phases
•
•
•
•
Bring cross functional teams
together using the models
Validate science merit functions
using models
Assess performance/quality risks
using the models
Use models to perform failure/fault
analysis
•
Testers (must now) validate the
models
•
Proactive integration (continuous)
•
Generate test plans/optimization
from the analysis models
•
Perform risk assessment based on
the models
•
Customers use model based
acquisition
•
Help with incremental integration
using models
Opportunities
How can MBSE Help?
• Improved, more rigorous communication with and across
disciplines including across analysis and IV&V teams.
• Support for continuous integration
• Better visualization and communication within domains by
SysML
• Improve opportunities for automating tasks across analysis and
IV&V
• Better linkage between system elements/interfaces and
requirements
• Improve coverage and gap analysis
Meeting Notes
I&V Scenario Breakout
Session
Challenges
• Duplication of effort – “we have
lots of models of the same
thing,” How do you know you
are testing the right thing?
• Model validation
• Consistency across lifecycle
• Define problem: cost, quality
and include Use cases
• Early V&V
• Consistent I&V Funding
• Early involvement of the
integration planners and
designers.
• Fidelity of model – keeping
models up to date with the latest
state of affairs
• Model management and
maintenance
• V&V folks not working in the
model
• Should create a model of the test
environment
Challenges
• IV&V fix problems from
early cycle
• Assumption of Vee
model? Alternatives are:
• Better communication &
coordination
• Spiral integration
• Formal methods
• Functional integration
• Cost avoidance
• Simulate, emulate, test
and validate
• Test environment a
development effort also
• Plug and play I&T
Challenges
• Things transferred between
the sides of the VEE
• Derive test cases from
systems model
• Push: UC/Requirements,
UC/CONOPS,
UC/architecture
• Concurrent Engineering
• Pull: What model influence
from IV&V
• Opportunities/Successes
• Minimize integration risk
through use of simulations
and emulations
• Capture dependencies
between each phase.
• No benchmarking
• Communication: better
communication standards e.g.
SysML
Challenges
• Methodology (process, methods and
tools)
• Method: Use of SysML BDDs
• High level activity diagrams
• Does SysML/architecture/etc help?
• Vee model: If modeling on the left
side, how do we pass on information
to the right side?
• Tools: manual -> more automation
only 50% being tested
• Knowledge transfer from tester to
tool/knowledge
• Model effort (Minimal) 50%
schedule reduction/avoidance
• Apply SysML satisfy links between
requirements and system model
• Coverage analysis/gap analysis:
e.g. Have all events passing
through a port been tested?
•
• Tying models together e.g. model
center
Model based design helps reduce
defects
Use Cases
• UC/Choose a development
model
• UC/introduce interdisciplinary
early
• UC/IV&V drivers – define main
topics, models: IV&V in charge?
• Models – integration plans
• UC/SE communication (Models as
enabler) – Bring team together
• UC/SE validate science merit
functions
• Lost trades/risk
• UC/Share models – constrained
by proprietary information
• UC/Subcontractor role –
Interfaces, performance, budget,
CONOPS
• UC/SE performance/quality
• Dynamic models
• UC/Domain knowledge experience
Use Cases
• UC/Fault – Failure analysis
• UC/Test planning and optimization
• UC/perform inter/intra model
consistency checking
• UC/Continuous integration support
• UC/Agile system development
• UC/USE MBE methods to solve
IV&V problems
• UC/Models used to help better
organize steps/artifacts in system
Vee
• UC/Proactive integration
• UC/New test role to validate the
model
• UC/Assess risks based on model
• UC/Model based acquisition
• UC/Help with incremental
validation
Opportunities
How can MBSE Help?
• Tying models together e.g. model center
• Communication: better communication
standards e.g. SysML
• Method: Use of SysML BDDs
• Methodology (process, methods and
tools)
• High level activity diagrams
• Does SysML/architecture/etc help?
• Model effort (Minimal) 50% schedule
reduction/avoidance
• Vee model: If modeling on the left side,
how do we pass on information to the
right side?
• Apply SysML satisfy links between
requirements and system model
• Tools: manual -> more automation only
50% being tested
• Coverage analysis/gap analysis: e.g.
Have all events passing through a port
been tested?
• Knowledge transfer from tester to
tool/knowledge
• Model based design helps reduce
defects
Opportunities
• IV&V fix problems from early cycle
• Better communication &
coordination
• Formal methods
• Cost avoidance
• Test environment a development
effort also
• Assumption of Vee model?
Alternatives are:
• Spiral integration
• Functional integration
• Simulate, emulate, test and
validate
• Plug and play I&T
• Things transferred between the
sides of the Vee
• Push: UC/Requirements,
UC/CONOPS, UC/architecture